|Publication number||US6898309 B2|
|Application number||US 10/799,213|
|Publication date||May 24, 2005|
|Filing date||Mar 12, 2004|
|Priority date||Apr 7, 2000|
|Also published as||US6754399, US7003178, US20010028736, US20040170317, US20050041854|
|Publication number||10799213, 799213, US 6898309 B2, US 6898309B2, US-B2-6898309, US6898309 B2, US6898309B2|
|Inventors||Daniel Pettigrew, François Paquin|
|Original Assignee||Autodesk Canada Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (17), Referenced by (11), Classifications (5), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application is a divisional application of Ser. No. 09/769,977, filed Jan. 25, 2001, now U.S. Pat. No. 6,754,399 for COLOR MATCHING IMAGE DATA by Daniel Pettigrew et al.,
1. Field of the Invention
The present invention relates to apparatus for processing image data, the method of processing image data and a computer-readable medium.
2. Description of the Related Art
Film editing and processing has historically been performed without the aid of digital image processing systems. In the past, the appearance of different film batches was controlled by a film colorist, whose job it is to ensure that when different film clips are spliced and edited together, their characteristics are not noticeably different. In order to achieve this, exposure times, negative types, and lighting conditions during a shoot have to be carefully controlled, and this is a highly specialised aspect of cinematography.
At the present time, increasing amounts of film editing and processing are being performed within a digital environment. The transfer of film processing to the digital domain has been relatively slow due to the enormous amounts of data that are required to be stored. Up to forty megabytes of data per frame of film is required in order to achieve equivalent film resolution in a digitised image. Slowly, however, as the cost of computer hardware and storage reduces, film processing is moving into the digital domain.
Once images have been digitised, it is possible to mix image data from a variety of formats, for example video mixed with film images, and to composite and edit these images together. It is known to imitate the film look by applying certain non-linear processing to video-sourced images, and so it may be possible to mix video and film images without the changeover being obvious between the two.
However, even though images can be freely mixed, composited and arranged in the digital domain, high quality editing still requires the same attention to detail that was performed by the film colorist. It is still necessary to control color matching between different image sources, whether they be from different reels of film, or from unrelated sources such as video. While it is widely known to employ controls such as contrast, color look-up tables, and various types of color manipulation, none of these methods provides an easy way of achieving a color match between images. A skilled colorist may be able to manipulate known systems in order to achieve a high degree of color matching. However, this remains a skilled and time consuming job. It is also a task that needs to be performed many times within a typical editing session.
According to an aspect of the present invention, there is provided apparatus for processing image data, including memory means for storing image data and processing instructions, and processing means for performing said image processing, wherein said instructions define operations to be performed by said processing means in order to process said image data, and said processing means is configured by said instructions such that: an input image is to be matched with a reference image, an initial output image has been generated from said input image and a matching transformation for processing image data has been initialised; and said operations include repeated steps of: identifying a first region in said output image and a reference image in said reference image, wherein said regions have a similar characteristic; comparing said regions to identify a difference transformation for said output image; updating said matching transformation by applying said difference transformation to it; and processing said input image with said matching transformation to update said output image.
A system for the processing of image data is illustrated in
A computer 103 facilitates the transfer of image data between the tape player 101 and the frame store 102. The computer 103 also facilitates the modification, processing and adjustment of image data to form an output clip that will eventually be stored onto digital tape. The computer is a Silicon Graphics Octane™. Images are previewed on a monitor 104 on which is also displayed a graphical user interface (GUI) to provide the user with several controls and interfaces for controlling the manipulation of image data. When processing image data, the user interacts with images and the graphical user interface displayed on the monitor 104 via a graphics tablet 105. For alphanumeric input, there is provided a keyboard 106, although facilities may be provided via the graphical user interface to facilitate occasional text input using the graphics tablet 105.
In addition to receiving image data from the tape player 101 and the frame store 102, the computer 103 may receive image and or other data over a network. The image processing system shown in
In a typical application, film clips are digitised and stored on digital tape for transfer to the system in FIG. 1. The film clips include several camera shots that are to be combined into the same scene. It is the task of the user or digital artist to combine and process this source image data into a single output clip that will be stored back onto tape for later transfer to film. Typical examples of this type of scene are where real images shot by a film camera are to be combined with artificially generated images and backgrounds, including scenes where actors are to be placed in computer-generated environments.
The computer 103 shown in
The switch 206 enables up to seven different non-blocking connections to be made between connected circuits. A graphics card 208 receives instructions from a CPU 201 or 202 in order to render image data and graphical user interface components on the monitor 104. A high bandwidth SCSI bridge 209 facilitates high bandwidth communications to be made with the digital tape player 101 and the frame store 102. An I/O bridge 210 provides input output interface circuitry for peripherals, including the graphics tablet 105, the keyboard 106 and a network. A second SCSI bridge 211 provides interface connections with an internal hard disk drive 212. This has a capacity of thirteen gigabytes. The second SCSI bridge 211 also provides connections to a CDROM drive 213, from which instructions for the central processing units 201 and 202 may be installed onto the hard disk 212.
Instructions for central processing units 201 and 202 determine how image data is processed. Several applications have been developed for image processing using processing hardware of the type shown in FIG. 2. The present applicant has image processing applications that include Flame™. The name Flame will henceforward refer to an improved version of Flame, operating in accordance with the present invention.
Steps performed by the user when operating the image processing system shown in
The contents of the main memory 207 shown in
In the present embodiment, the main application consists of instructions 402 for Flame. These instructions 402 include color matching instructions 403, along with other instructions for processing image data. Main areas of memory used by the color matching instructions include input image data 404, reference image data 405 and output image data 406. Additionally, the main memory includes workspace 407, which may provide dynamically allocatable regions of memory as required by the execution of instructions 401, 402 and 403.
Steps performed during image processing 304 are illustrated in FIG. 5. At step 501 the user instructs the image processing system to import images from a film clip stored on a digital tape in the tape player 101. The clip comprises a plurality of images, with film images typically being stored as a sequence of twenty four frames of image data for each second of playback time. The clip images are transferred, with formatting if necessary, into the frame store 102. Once in the frame store 102, the image data can be manipulated at high speed, including preview and playback in real time. At step 502 the user compares the clip imported at step 501 with a reference clip. The reference clip is a clip comprising image data with a desired set of color characteristics.
Typically, a reference clip consists of a clip taken from a single camera, in a single take. Thus, all the characteristics of film chemistry and lighting are substantially consistent from frame to frame, across the entire duration of the reference clip. When editing together several clips for a single image sequence, differences in image color characteristics, as well as brightness and contrast, will be noticeable. This is particularly true if the component clips come from completely different sources, for example, when the same scene has been shot several times over a period of several days or weeks, or when combining together clips for a single scene that have been shot in different studios. It is also possible that some of the clips are sourced from video cameras.
In order to ensure that the single resulting clip does not have abrupt changes in aesthetics, it is the job of a film colorist to control a color matching process, perhaps by modifying the exposure times of film negatives during subsequent development and chemical processing. With the advent of digital processing, it has theoretically become possible to modify image color characteristics in an arbitrary manner. However, this requires the highly skilled intervention of a film colorist, who is able to determine what color modifications are required.
Thus, at step 502, the user compares color characteristics of a reference clip with the imported clip. At step 503 a question is asked as to whether color matching is required. If not, control is directed to step 508. Alternatively, control is directed to step 504. At step 504 it is known that the clip imported at step 501 requires color matching with the reference clip. In order to perform the color matching procedure, the user first identifies a single input image from the imported clip. This input image is a single frame of image data, and is stored at 404 in main memory 207. At step 505, the user identifies a single reference image from the reference clip, stored at 405 in main memory. The user identifies the input 404 and the reference image 405, intending that the identified input image is to be matched with the reference image, and thereafter, if the images 404 and 405 have been appropriately identified by the user, it may be possible to apply the same color matching transformation to all images in an input clip.
At step 506, the color transformation required for matching the input image 404 to the reference image 405 is defined. At step 507 the color transformation defined at step 506 is applied to the remaining images in the imported clip. At step 508 a question is asked as to whether it is required to import another film clip. If so, control is directed to step 501. Alternatively, if all input clips have been imported, operations may then proceed to step 509, where the input clips are processed and combined in various ways, in accordance with the known arts of digital image compositing and processing.
The step of defining a color transformation 506 shown in
At step 601 a matrix M is created and initialised as an identity matrix. Input image data 404 comprises pixels, each of which have a red, green and blue (RGB) color value. Matrix M will eventually be used to modify RGB values so as to perform color matching. However, at step 601 matrix M is initialised such that, when applied to an image RGB pixel, the resulting RGB values are unchanged. At step 602 the input image 404 is copied to an output image 406. Throughout color matching, the input image 404 will remain unchanged, and an output image 406 will be generated in response to matrix M at each iteration of the procedure, in order to monitor and facilitate further refinement of the color matching transformation.
At step 603 the user identifies a pair of comparable regions in the output image 406 and reference 405 image. In order to perform color matching successfully, it is necessary to perform adjustments to color characteristics across a range of luminance values. Thus, in selecting a pair of regions at step 603, the user identifies regions having one of three possible characteristics: highlight, shadow or overall.
Initially, the user may identify highlight regions. Thus, a region in the output image having a high brightness is identified. Then, a region in the reference image is also identified, which contains the correct color and or luminance characteristics for the region selected in the output image. The type of identified region is characterised at step 604.
The pair of regions identified at step 603 defines a difference in color and or luminance characteristics. At step 605, differences between the identified regions are analysed in order to define a difference matrix or matrices. These matrices are characterised in that, if applied to the image data of the input image 404, the identified region in the output image has its color and or luminance characteristics matched to those of the identified reference region. Other parts of the output image would also be affected, to varying degrees. The hope is that these results will also be appropriate, and have the effect of bringing the overall color and or luminance characteristics of the whole output image 406 closer to that of the reference image 405. The type of processing performed at step 605 is dependent upon the region characteristic identified at step 604.
It is intended to modify the matrix many times by repeating steps 603, 604 and 605. The resulting difference matrix would only represent the latest incremental difference, and not the color difference between the input image 404 and the reference image 405. It is desirable to define this full difference transformation, so that it may then be subsequently applied to several frames of image data from the originally imported film clip. Furthermore, an improvement in processing quality may result in certain implementations where image pixel RGB values are restricted in precision, if only a single transformation is applied to the input image 404.
In order to define the full transformation, at step 606 the difference matrix or matrices defined at step 605 are concatenated onto matrix M. In matrix arithmetic, this is written as M=MSR, where S and R are differential transformations defined as a result of the analysis performed at step 605. The resulting matrix M is applied to the input image 404, to create new data for the output image 406. At step 608, a question is asked as to whether the result has achieved a close enough color match between the output and reference images, so that this transformation may now be applied. If so, this concludes the steps for defining a color matching transformation 506.
On a subsequent iteration, at step 603 the user identifies a region in the output image that is dark. The user selects a corresponding dark region in the reference image 404 towards which the characteristics of the output region should be transformed. The reference region is characterised as a shadow region at step 604. At step 605 difference matrices are defined, and these are concatenated onto matrix M. Again, the output image 406 is updated at step 607.
In another iteration, the user identifies a pair of regions having overall characteristics. In other words, a region that includes a range of luminance values characteristic of the whole image, or important parts of it. The processing performed with respect to an overall region is different at step 605, where only a single difference transformation matrix is defined. This is concatenated onto matrix M at step 607.
Having defined highlight, shadow and overall region transformations, resulting in a matrix M that encompasses these, it is possible for the user to continue selecting and matching highlight, shadow or overall regions, in any order, until the output image is sufficiently matched in appearance with the reference image 405. Eventually, at step 608, the user will decide to apply the resulting matrix M to a frame or frames of input image data in the imported clip.
The steps shown in
The user interface presented to the user, for implementing the steps of
A select button 702 instructs the image processing system to permit the user to identify regions in the output 406 and reference images 405 on the display. The user drags the cursor over an area of each image in order to identify a rectangular region of pixels that will be used for subsequent analysis. An output region 703 and a reference region 704 are shown in the example in FIG. 7. The user selects the output region 703 first, and the reference region second. As soon as the user ends the dragging operation to define the second region 704, processing takes place, as defined at steps 605, 606 and 607 in FIG. 6. The user may define a different reference region 704 without defining a new output region 703. Each time the user completes the drag operation to select the reference region, processing is activated automatically. This enables the user to obtain a useful and accurate result very quickly.
A reset button 706 enables the user to cancel all concatenated transformations, and to restart the steps shown in FIG. 6. An apply button 707 enables the user to apply the transformation, that has been constructed over several iterations, to multiple image data. Save and load buttons 708 and 709 enable the user to save the transformation for subsequent use in a different session. An indication of the type of region identified at step 604 is shown at 710, so that the user is aware of the type of matching that will occur when processing commences after a new reference region 704 is defined.
Details of the characterising process performed at step 604 in
Alternatively, if all identified pixels have been characterised, control is directed to step 901 in FIG. 9. At step 901 SHADOW is divided by NPIXELS, the number of pixels in the identified region 704. At step 902 a question is asked as to whether SHADOW is greater than 0.66. If so, the region 704 is characterised as a shadow region at step 903. Alternatively, at step 904 MID-TONE is divided by NPIXELS, and at step 905 a question is asked as to whether MID-TONE is greater than 0.66. If so, the region 704 is characterised as a highlight region at step 908. Alternatively, at step 909, if the region is neither a shadow nor a highlight region, the region is characterised as an overall color region. This concludes the steps performed within step 604, in which the identified reference region is characterised. Characterisation of the output region 703 is not performed, as this is a region that is intended for transformation, and hence not reliable as a source of characterisation data.
Once characterisation has been performed at step 604 in
In the invention it is considered useful to define modifications to pixel data, in such a way that it is possible to transform a pixel's luminance without varying its color, or, as an alternative, to transform a pixel's color without transforming its luminance. By characterising differences between identified regions 703 and 704 in separate terms of luminance and color, it becomes possible to perform general shifts in luminance and then color that are appropriate to a range of color values in an image.
Y′PbPr color space may be considered as having a cylindrical shape with a central axis Y′, that is a vector extending out from the origin of RGB color space, as shown at 1101. Conversion between these color spaces may be achieved by a matrix, and the parameters required for a transformation from RGB to Y′PbPr are detailed at 1102. Transformation from RGB to Y′PbPr may be assigned to a matrix A. The inverse of A, A−1, provides transformation from Y′PbPr to RGB. There is an intuitive relationship between these color spaces for colors of pure black and pure white, as shown at the bottom of FIG. 11.
Y′PbPr color space may be viewed as shown in FIG. 12. The projection shown at 1201 ignores the luminance dimension, and permits characterisation of pure color. Shades of grey from black to white are located at the origin 1202. The amount of color can be quantified by a saturation value, which can be viewed geometrically as the length of the vector from the centre 1202 to a color's PbPr co-ordinates. All three dimensions of Y′PbPr color space may be viewed as shown at 1204.
Details of highlight processing 1002 are shown in FIG. 13.
At step 1303 a scaling matrix S is identified that, when applied to RGB pixels having luminance Yo, translates their RGB values such that their luminance becomes Yr. By considering luminance in isolation, color values are not changed by this matrix. RGB values having a luminance other than Yo have their luminance translated proportionately, again without affecting color. Thus, steps 1302 and 1303 result in a scaling matrix S that modifies the luminance of pixels in order to match the luminance of output region 703 to that of reference region 704, without changing the color of any pixel.
At step 1304 matrix S is applied to RoGoBo and RrGrBr, prior to color processing in response to their modified values.
At step 1305, Pb and Pr color co-ordinates are obtained from the updated color values created at step 1304. This is done using the Pb and Pr equations shown at 1102 in FIG. 11. For the output region 703, values Pbo and Pro are created. For the reference region 704, values Pbr and Prr are created.
At step 1306 a sheer matrix R is identified which, when applied to RGB pixels having color (Pbo,Pro) translates their RGB values such that their color becomes (Pbr,Prr). By considering color in isolation from luminance, a pure color shift is achieved. Pixels at different color values from (Pbo,Pro) have their color modified proportionately, again without changing luminance.
Steps 1304, 1305 and 1306 result in a sheer matrix R that modifies RGB pixel values in order to match the color of output region 703 to that of reference region 704.
A graphical illustration of highlight processing, performed in accordance with the steps of
The sheer requirement for defining matrix R is illustrated at 1402 and 1403, for Pb and Pr dimensions respectively. In highlight processing 1002, the black end of Y′PbPr color space is fixed, and the sheer is applied to the white end, in such a way that Pbo is moved to Pbr, and Pro is moved to Prr. Other points in color space are modified proportionately.
Shadow processing 1003 is illustrated in FIG. 15. In shadow processing, all transformations are performed with the white end of Y′PbPr color space fixed, and movements occurring at the black end. Thus, at 1501, the scaling matrix S that is defined results in point of luminance Yo being moved to the same color at luminance Yr. However, unlike highlight processing, this scaling is performed with the white end fixed. As will be understood by those knowledgeable in the art of geometric matrix manipulations, this is achieved by a concatenation of three operations: a translation, a scaling, and a re-translation to undo the effect of the translation. Matrix S is a concatenation of these three operations.
The sheer operation performed in shadow processing 1003 is illustrated at 1502 and 1503, with respect to Pb and Pr dimensions of Y′PbPr color space. As with highlight processing, a sheer is applied to an end of Y′PbPr color space such that Pbo moves to Pbr, and Pro moves to Prr. However, the sheer matrix R encompasses operations such that the sheer is applied to the dark end of Y′PbPr color space, while the light end remains fixed.
Matrices S and R are defined using known three-dimensional mathematical techniques and equations for matrices. In particular, in order to define a geometric manipulation in Y′PbPr color space, for application to pixels having RGB values, it is necessary to translate the pixel RGB value to Y′PbPr, apply a matrix s operating in Y′PbPr color space, and then convert back to RGB, by performing the operation:
in which A−1 is the inverse of matrix A, which translates from RGB to Y′PbPr color space, as indicated in FIG. 11. The same type of operation may be achieved for the sheer matrix R:
thereby simplifying the concatenation of matrix M, that can operate directly on pixels in RGB color space:
However, the geometric transformations encoded within matrix M ensure that luminance and color translations and sheers are applied independently and appropriately.
The third type of region that may be selected by the user at step 603 is an overall region. Pixels in this type of region have a wide range of luminance values, and in the corresponding matching process, luminance is ignored. The overall color spread is compared and matched. Overall color processing 1004 is summarised in FIG. 16. At step 1601 the maximum saturation, So, of the output region 703 is calculated. At step 1602 the maximum saturation, Sr, of the reference region 704 is calculated. At step 1603 a scaling matrix T is defined that, when applied to RGB pixels having saturation So, results in RGB values having saturation Sr.
Processing for the identification of maximum saturation values, required at steps 1601 and 1602, is detailed in FIG. 17. At step 1701 a variable SMAX is initialised to zero. At step 1702 the first or next pixel is selected. At step 1703 values for Pb and Pr are obtained from the RGB values, according to the equations given at 1102 in FIG. 11.
At step 1704 the magnitude of the saturation vector is calculated as being equal to the square root of the sum of the squares of Pb and Pr for the selected pixel. At step 1705 a question is asked as to whether the value calculated at step 1704 is greater than SMAX. If so, SMAX is updated with this value at step 1706. Alternatively, control is directed to step 1707. At step 1707 a question is asked as to whether there are any other pixels in region 703 or 704 that require processing. If so, control is directed to step 1702. Alternatively, this completes the calculations for the region, with the maximum color saturation given by the current value of variable SMAX.
An illustration of the overall matching process is shown in FIG. 18. Point 703 represents the pixel in output region 703 that has the greatest saturation. Point 704 represents the pixel in reference region 704 that has the greatest saturation. A matrix T is defined such that the saturation So is made equal to saturation Sr. However, only the magnitude of the vector So changes, not its direction. The direction of the vector may change subsequently, as a result of sheer operations being applied to one or the other end of Y′PbPr color space, in accordance with highlight and or shadow processing. Again, a matrix t in Y′PbPr color space may be defined, and converted for use on RGB data by the relation:
and matrix M is updated t step 606 in accordance with:
Thereafter RGB pixels may be processed by application of matrix M, and only a single matrix vector product is required per pixel, even though the transformations defined and concatenated into matrix M over several iterations may be complex. Once a satisfactory color match has been achieved between the output image 406 and the reference image 405, matrix M may finally be applied to multiple images in a film clip, or alternatively just the single frame of input image data 404, as required. The invention is applicable to image data of any source or format, where color matching between images is required.
It is possible for the color characteristic of the input clip or the reference clip to change gradually over a sequence of frames. Matrix M may be animated to take into account these variations. For example, a first matrix M1 is defined for frame one of a clip, and a second matrix M10 is defined for frame ten of a clip. Interpolation is performed between these two matrices for the intermediate frames two to nine, resulting in the matrices M2 to M9 being applied to these intermediate frames. This may be considered a case of animating the matrix M in response to color changes occurring in the film or other source material.
An animation curve may be defined by the user for controlling interpolation between the matrices of different frames. In the above example, the animation curve would modify the rate of change from M1 to M10. This may take into account sudden variability, for example when clouds obscure sunlight at frame seven. The animation curve may be defined by user-defined control points along a spline, or other curve.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5559604 *||Sep 27, 1994||Sep 24, 1996||Toyo Ink Manufacturing Co., Ltd.||Color matching apparatus for reproducing the same color under different illuminants|
|US5579031||Feb 14, 1995||Nov 26, 1996||E. I. Du Pont De Nemours And Company||Color matching method and apparatus|
|US5619596||Oct 6, 1993||Apr 8, 1997||Seiko Instruments Inc.||Method and apparatus for optical pattern recognition|
|US5724450 *||Sep 30, 1994||Mar 3, 1998||Apple Computer, Inc.||Method and system for color image compression in conjunction with color transformation techniques|
|US5727093||Feb 6, 1996||Mar 10, 1998||Canon Kabushiki Kaisha||Image processing method and apparatus therefor|
|US6046723 *||Nov 17, 1997||Apr 4, 2000||Eastman Kodak Company||Method and apparatus for achieving color-appearance matching for an image viewed in surrounds of different relative luminances|
|US6075905||Jul 16, 1997||Jun 13, 2000||Sarnoff Corporation||Method and apparatus for mosaic image construction|
|US6111984||Jan 21, 1998||Aug 29, 2000||Fujitsu Limited||Method for matching input image with reference image, apparatus for the same, and storage medium storing program for implementing the method|
|US6188800 *||Aug 27, 1997||Feb 13, 2001||Yamaha Corporation||Two-dimensional spatial transformation system for video processing|
|US6504957||Feb 27, 2002||Jan 7, 2003||General Electric Company||Method and apparatus for image registration|
|US6559982 *||Jan 12, 1999||May 6, 2003||Hewlett-Packard Company||Accurate monitor to printer color reproduction technique|
|US6563945||Jan 29, 2001||May 13, 2003||Jack M. Holm||Pictorial digital image processing incorporating image and output device modifications|
|US6625216||Jan 27, 1999||Sep 23, 2003||Matsushita Electic Industrial Co., Ltd.||Motion estimation using orthogonal transform-domain block matching|
|US6628823||Feb 21, 2003||Sep 30, 2003||Jack M. Holm||Pictorial digital image processing incorporating adjustments to compensate for dynamic range differences|
|US6766055 *||May 20, 2002||Jul 20, 2004||Canon Kabushiki Kaisha||Method of extracting image from input image using reference image|
|GB2157529A||Title not available|
|WO1998010586A1||Sep 5, 1997||Mar 12, 1998||Discreet Logic Inc.||Image processing system converting between video data and addressed data|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7072510 *||Aug 28, 2002||Jul 4, 2006||Autodesk Canada Co.||Adjusting data representing image pixel color|
|US7693341||Apr 21, 2006||Apr 6, 2010||Apple Inc.||Workflows for color correcting images|
|US8022964||Apr 21, 2006||Sep 20, 2011||Apple Inc.||3D histogram and other user interface elements for color correcting images|
|US8031962||Apr 1, 2010||Oct 4, 2011||Apple Inc.||Workflows for color correcting images|
|US8203571||Sep 7, 2011||Jun 19, 2012||Apple Inc.||3D histogram for color images|
|US9390153 *||Feb 21, 2012||Jul 12, 2016||Dmitriy Tochilnik||User-configurable radiological data transformation routing and archiving engine|
|US20030206665 *||Aug 28, 2002||Nov 6, 2003||Autodesk Canada Inc.||Processing image data|
|US20070247475 *||Apr 21, 2006||Oct 25, 2007||Daniel Pettigrew||3D histogram and other user interface elements for color correcting images|
|US20070247647 *||Apr 21, 2006||Oct 25, 2007||Daniel Pettigrew||3D lut techniques for color correcting images|
|US20070247679 *||Apr 21, 2006||Oct 25, 2007||Daniel Pettigrew||Workflows for color correcting images|
|US20100188415 *||Apr 1, 2010||Jul 29, 2010||Apple Inc.||Workflows for Color Correcting Images|
|U.S. Classification||382/162, 382/278|
|Nov 24, 2008||FPAY||Fee payment|
Year of fee payment: 4
|Mar 24, 2009||AS||Assignment|
Owner name: AUTODESK, INC., CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:AUTODESK CANADA CO.;REEL/FRAME:022445/0222
Effective date: 20090225
Owner name: AUTODESK, INC.,CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:AUTODESK CANADA CO.;REEL/FRAME:022445/0222
Effective date: 20090225
|Dec 12, 2012||FPAY||Fee payment|
Year of fee payment: 8
|Dec 12, 2012||SULP||Surcharge for late payment|
Year of fee payment: 7
|Dec 15, 2016||FPAY||Fee payment|
Year of fee payment: 12
|Dec 15, 2016||SULP||Surcharge for late payment|
Year of fee payment: 11